Examples: Modifying Tabular Output by Using the TEMPLATE Procedure | SAS 9.1 Output Delivery System Users Guide

Example 1: Editing a Table Definition that a SAS Procedure Uses

PROC TEMPLATE features:

EDIT statement
Header attributes
- JUST=
- STYLE=

Table attributes
- DOUBLE_SPACE=
- OVERLINE=
- UNDERLINE=

Other ODS features:

ODS HTML statement
ODS SELECT statement

Data set: STATEPOP

Program Description

Note: This example uses filenames that might not be valid in all operating environments. To successfully run the example in your operating environment, you might need to change the file specifications. See Appendix 3, 'ODS HTML Statements for Running Examples in Different Operating Environments,' on page 649.

This example customizes the table definition for the Moments output object from PROC UNIVARIATE. The first program uses the table definition that SAS supplies to generate both Listing output and HTML output of the Moments object.

The second program

creates and edits a copy of the default table definition.
edits a header within the table definition.
sets column attributes to enhance the appearance of both the HTML and the Listing output.

Program 1: Using the Default Table Definition that SAS Provides

Set the SAS system options. The OPTIONS statement controls several aspects of the Listing output. None of these options affects the HTML output.

 options nodate pageno=1 pagesize=60 linesize=72;

Create the HTML output and specify the name of the HTML file. The ODS HTML statement opens the HTML destination and creates HTML output. It sends all output objects to the external file defaultmoments-body.htm in the current directory. Some browsers require an extension of .htm or .html on the filename.

 ods html body='defaultmoments-body.htm';

Select the output objects for the report. The ODS SELECT statement sends one output object, Moments , to the open ODS destinations. Both the Listing and the HTML destinations are open . (To learn the names of the output objects, run the procedure with the ODS TRACE ON statement in effect. See 'Example' on page 200.)

 ods select moments;

Compute the descriptive statistics for one variable. PROC UNIVARIATE computes the univariate statistics for one variable, CityPop_90. It uses the default table definition, base.univariate.moments from the template store sashelp.tmplmst .

 proc univariate data=statepop mu0=3.5;     var citypop_90;  title 'Default Moments Table';  run;

Stop the creation of the HTML output. The ODS HTML statement closes the HTML destination and all the files that are associated with it. You must close the destination before you can view the output with a browser.

 ods html close;

Default Listing Output

Output 10.3: Listing Output from PROC UNIVARIATE (Default Moments Table)

 Default Moments Table                           1                       The UNIVARIATE Procedure      Variable:   CityPop_90   (1990 metropolitan pop in millions)                                Moments  N                            51    Sum Weights                51  Mean                 3.87701961    Sum Observations      197.728  Std Deviation        5.16465302    Variance           26.6736408  Skewness             2.87109259    Kurtosis            10.537867  Uncorrected SS       2100.27737    Corrected SS       1333.68204  Coeff Variation       133.21194    Std Error Mean     0.72319608

HTML Output from PROC UNIVARIATE (Default Moments Table)

Display 10.5: Default HTML Output (Viewed with Microsoft Internet Explorer)

Program 2: Using a Customized Table Definition

Specify the search path in order to locate the table definition. The ODS PATH statement specifies which locations to search for definitions that were created by PROC TEMPLATE, as well as the order in which to search for them. The statement is included to ensure that the example works correctly. However, if you have not changed the path, you do not need to include this statement because it specifies the default path.

 ods path sasuser.templat(update) sashelp.tmplmst(read);

Create a modified table definition base.univariate.moments . The EDIT statement looks in the available template stores for a table definition called base.univariate.moments .By default, it first looks in SASUSER.TEMPLAT, but it finds nothing. Next , it looks in SASHELP.TMPLMST, which contains the table definitions that SAS provides. Because the EDIT statement can read this definition, this is the one that it uses. The program does not specify a destination for the edited definition, so PROC TEMPLATE writes to the first template store in the path that it can write to, which is SASUSER.TEMPLAT. Therefore, it creates a table definition of the same name as the original one in SASUSER.TEMPLAT. (See 'ODS PATH Statement' on page 149).

(To learn the name of the table definition that a procedure uses, run the procedure with the ODS TRACE ON statement in effect. See 'Example' on page 200).

 proc template;     edit base.univariate.moments;

Specify changes to the Moments output object. These three table attributes affect the presentation of the Moments output object in the Listing output. They have no effect on its presentation in the HTML output. DOUBLE_SPACE= double spaces between the rows of the output object. OVERLINE= and UNDERLINE= draw a continuous line before the first row of the table and after the last row of the table.

 double_space=on;  underline=on;  overline=on;

Modify a table element. This EDIT statement edits the table element head within the table definition.

 edit head;

Modify the appearance of the header. The STYLE= attribute alters the style element that produces the head table element. The style element header is defined in the default style definition, styles.default . Many procedures, including PROC UNIVARIATE, use this style element to produce headers for tables and columns . (For information on viewing a style definition, see

'What Style Definitions Are Shipped with SAS Software?' on page 30.) In this case, the STYLE= attribute specifies green for the foreground color and italic for the font style. All other attributes that are included in header remain in effect. The STYLE= attribute affects only the HTML output.

 style=header{foreground=green font_style=italic};

Left justify the header text. The JUST= attribute left-justifies the text of the header in both the Listing and the HTML output.

 just=left;

Stop the editing of the table element and the table definition. The first END statement ends the editing of the table element head . The second END statement ends the editing of the table base.univariate.moments .

 end;     end;  run;

Create the HTML output and specify the name of the HTML file. The ODS HTML statement opens the HTML destination and creates HTML output. It sends all output objects to the external file custommoments-body.htm in the current directory. Some browsers require an extension of .htm or .html on the filename.

 ods html body='custommoments-body.htm';

Select the output objects for the report. The ODS SELECT statement sends one output object, Moments , to the open ODS destinations. Both the Listing and the HTML destinations are open. (To learn the names of the output objects, run the procedure with the ODS TRACE ON statement in effect. See 'Example' on page 200.)

 ods select moments;

Compute the descriptive statistics for one variable. PROC UNIVARIATE computes the univariate statistics for one variable, CityPop_90. This is the same PROC UNIVARIATE step that was used in 'Program 1: Using the Default Table Definition that SAS Provides' on page 516. The actual results of the procedure step are the same in this case, but they are presented differently because the procedure uses the edited table definition. It does so because when it looks for base.univariate.moments , it looks in the first template store in the path, SASUSER.TEMPLAT. If you wanted to use the table definition that is supplied by SAS, you would have to change the path with the ODS PATH statement (see 'ODS PATH Statement' on page 149).

 proc univariate data=statepop mu0=3.5;     var citypop_90;  title 'Custom Moments Table';  run;

 ods html close;

Customized Listing Output

Output 10.4: Listing Output (Customized Moments Table) from PROC UNIVARIATE

 Custom Moments Table                           1                       The UNIVARIATE Procedure     Variable:   CityPop_90   (1990 metropolitan pop in millions)  Moments  ----------------------------------------------------------------- N                           51      Sum Weights                51  Mean                3.87701961      Sum Observations      197.728  Std Deviation       5.16465302      Variance           26.6736408  Skewness            2.87109259      Kurtosis            10.537867  Uncorrected SS      2100.27737      Corrected SS       1333.68204  Coeff Variation      133.21194      Std Error Mean     0.72319608  -----------------------------------------------------------------

Customized HTML Output

Display 10.6: Customized HTML Output (Customized Moments Table) from PROC UNIVARIATE (Viewed with Microsoft Internet Explorer)

Example 2: Comparing the EDIT Statement with the DEFINE TABLE Statement

PROC TEMPLATE features:

EDIT statement
COLUMN statement
DEFINE statement
- STYLE= attribute

NOTES statement
DYNAMIC statement

Other ODS features:

ODS PATH statement
ODS HTML statement
ODS HTML CLOSE statement

Program Description

This example compares the use of an EDIT statement with a DEFINE TABLE statement for the same table definition. The first program uses the EDIT statement to change the Base.Summary table definition. The foreground color of the NOBS column is changed to green. The other definitions and attributes of the Base.Summary table definition remain the same. The second program uses the DEFINE TABLE statement to define a new table using the same name, Base.Summary . The NOBS column is the only column defined in the new table definition. When the PROC SUMMARY step executes, only the NOBS column is printed. The only style attribute that is used to format the column is the foreground=green attribute.

Program 1

Edit the existing table definition Base.Summary . The ODS PATH statement specifies which item store to search first for the table definition. The EDIT statement edits the table definition Base.Summary . The modified table definition Base.Summary is written to the WORK.TEMPLAT item store.

 ods path work.templat (update) sashelp.tmplmst (read);  proc template;      edit Base.Summary;          edit nobs;          style={foreground=green};      end;     end;  run;  ods html file="temp.html";  proc summary data=sashelp.class print;  class age;  run;  ods html close;

The column labeled AGE remains in the output because AGE is defined as a dynamic variable which is passed to the original Base.Summary table definition and AGE is specified as the CLASS variable. The attributes of the NOBS column are modified in the EDIT statement where the NOBS column is defined.

Display 10.7: HTML Output Using an Edited Table Definition for Base.Summary

Output 10.5: Base.Summary Table Definition Modified by the EDIT Statement

The modified Base.Summary table definition changes the foreground color of the NOBS column to green. The vertical alignment and heading of the NOBS column, and the other table attributes, are retained from the default table definition and stay the same. To view the Base.Summary table definition created by Program 1, follow these steps.

Submit the following command in the command bar:
```
  odstemplates  
```
Double-click the item store WORK.TEMPLAT .
Double-click the item store Base .
Right-click the table definition Summary and select Open . The table definition Base.Summary is displayed in the Template Browser window.

 proc template;     define table Base.Summary / store = SASUSER.TEMPLAT;        notes "Summary table for MEANS and SUMMARY";        dynamic one_var one_var_label one_var_name clmpct;        column class nobs id type ways (varname) (label) (min) (max) (range) (n)           (nmiss) (sumwgt) (sum) (mean) (uss) (css) (var) (stddev) (cv) (stderr           ) (t) (probt) (lclm) (uclm) (skew) (kurt) (median) (mode) (q1) (q3) (           qrange) (p1) (p5) (p10) (p25) (p50) (p75) (p90) (p95) (p99);        header h;        define p99;           header = "99th Pctl";           generic;        end;        define p95;           header = "95th Pctl";           generic;        end;        define p90;           header = "90th Pctl";           generic;        end;        define p75;           header = "75th Pctl";           generic;        end;        define p50;           header = "50th Pctl";           generic;        end;        define p25;           header = "25th Pctl";           generic;        end;        define p10;           header = "10th Pctl";           generic;        end;        define p5;           header = "5th Ptcl";           generic;        end;        define p1;           header = "1st Pctl";           generic;        end;        define qrange;           header = "Quartile Range";           generic;        end;        define q3;           header = "Upper Quartile";           generic;        end;        define q1;           header = "Lower Quartile";           generic;        end;  define mode;          header = "Mode";          generic;        end;         define median;           header = "Median";           generic;        end;         define kurt;           header = "Kurtosis";           generic;        end;        define skew;           header = "Skewness";           generic;       end;       define uclm;          define header huclm;             text "Upper " clmpct BEST8. %nrstr("%%/CL for Mean");             split = "/";          end;          header = huclm;          generic;       end;       define lclm;          define header hlclm;             text "Lower " clmpct BEST8. %nrstr("%%/CL for Mean");             split = "/";          end;          header = hlclm;          generic;       end;       define probt;          parent = Common.ParameterEstimates.Probt;          generic;       end;       define t;          parent = Common.ParameterEstimates.tValue;          generic;       end;       define stderr;          header = "Std Error";          parent = Common.ParameterEstimates.StdErr;          generic;       end;       define cv;          header = "Coeff of Variation";          generic;       end;       define stddev;          header = "Std Dev";          generic;       end;       define var;          header = "Variance";          generic;       end;       define css;          define header hcss;             text2 "CSS";             text "Corrected SS";          end;          header = hcss;          generic;       end;       define uss;          define header huss;             text2 "USS";             text "Uncorrected SS";          end;          header = huss;          generic;       end;     define mean;           header = "Mean";           generic;        end;        define sum;           header = "Sum";           generic;        end;        define sumwgt;           header = "Sum Wgts";           generic;        end;        define nmiss;           header = "N Miss";           generic;        end;        define n;           header = "N";           generic;        end;        define range;           header = "Range";           generic;        end;        define max;           define header hmax;              text2 "Max";              text "Maximum";           end;           header = hmax;           generic;        end;        define min;           define header hmin;              text2 "Min";              text "Minimum";           end;           header = hmin;           generic;        end;        define label;           header = "Label";           id;           generic;        end;        define varname;           header = "Variable";           id;           generic;        end;        define ways;           header = "Ways";           vjust = T;           id;        end;        define type;           header = "Type";           vjust = T;           id;        end;        define id;           vjust = T;           id;           generic;        end;  define nobs;           header = "N Obs";           vjust = T;           style = {              foreground = green              };              id;           end;  define class;           vjust = T;           id;           generic;           blank_internal_dups;        end;        define h;           text "Analysis Variable : " one_var_name " " one_var_label;           space = 1;           just = C;           print = one_var;           spill_margin;        end;        required_space = 5;        underline;        overline;        byline;        use_format_defaults;        double_space;        split_stack;        use_name;        order_data;        classlevels;     end;  run;

Program 2

Define the table Base.Summary . The ODS PATH statement specifies which item store to search first for the table definition. The DEFINE TABLE statement creates a new table definition Base.Summary . The new table definition Base.Summary is written to the WORK.TEMPLAT item store.

 ods path work.templat (update) sashelp.tmplmst (read);  proc template;     define table Base.Summary;        notes "Summary table for MEANS and SUMMARY";        dynamic clmpct one_var_name one_var_label one_var;        column class nobs id type ways (varname) (label) (min) (max) (range) (n           ) (nmiss) (sumwgt) (sum) (mean) (uss) (css) (var) (stddev) (cv) (           stderr) (t) (probt) (lclm) (uclm) (skew) (kurt) (median) (mode) (q1)           (q3) (qrange) (p1) (p5) (p10) (p25) (p50) (p75) (p90) (p95) (p99);        define nobs;          style={foreground=green};        end;     end;  run;  ods html file="temp.html";  proc summary data=sashelp.class print;  class age;  run;  ods html close;

The column labeled AGE is missing because it was not defined in the new table definition Base.Summary . The new table definition only defined the NOBS column with a green foreground and no column headings.

Display 10.8: HTML Output that Uses the Table Definition Base.Summary.

Output 10.6: Base.Summary Table Definition Created by the DEFINE TABLE Statement

The Base.Summary table definition defines the foreground color of the NOBS column to green. Because the vertical alignment and heading of the NOBS column, and the other table attributes, are not defined, they are no longer part of the Base.Summary table definition. To view the table definition Base.Summary created by Program 2, follow these steps.

Submit the following command: odstemplates
Double-click the item store WORK.TEMPLAT .
Double-click the item store Base .
Right-click the table definition Summary and select Open . The table definition Base.Summary is displayed in the Template Browser window.

 proc template;     define table Base.Summary / store = WORK.TEMPLAT;        notes "Summary table for MEANS and SUMMARY";        dynamic clmpct one_var_name one_var_label one_var;        column class nobs id type ways (varname) (label) (min)        (max) (range) (n)(nmiss) (sumwgt) (sum) (mean) (uss) (css)        (var) (stddev) (cv) (stderr) (t) (probt) (lclm) (uclm) (skew)        (kurt) (median) (mode) (q1) (q3) (qrange) (p1) (p5) (p10)        (p25) (p50) (p75) (p90) (p95) (p99);        define nobs;           style = {              foreground = green              };        end;     end;  run;

Example 3: Creating a New Table Definition

PROC TEMPLATE features:

Table attributes:
- DOUBLE_SPACE=
- OVERLINE=
- UNDERLINE=

DEFINE TABLE statement:
- COLUMN statement
- DEFINE statement (for columns)
  - GENERIC= attribute
  - HEADER= attribute
  - ID= attribute
  - STYLE= attribute
  - VJUST= attribute
- DEFINE statement (for headers)
  - TEXT statement
  - STYLE= attribute
  - SPACE= attribute
- DEFINE FOOTER statement

HEADER statement
MVAR statement

Other ODS features:

ODS HTML statement
FILE statement with ODS= option
PUT statement with _ODS_ argument

Program Description

This example creates a custom table definition for an output data set that PROC MEANS produces.

Program 1: Producing an Output Data Set with PROC MEANS

Set the SAS system options. The OPTIONS statement controls several aspects of the Listing output. None of these options affects the HTML output.

 options nodate pageno=1 pagesize=60 linesize=72;

Create formats for the variables Year and School . PROC FORMAT creates formats for Year and School.

 proc format;     value yrFmt . = " All";     value $schFmt ' ' = "All       ";     run;

The data set Charity contains information about students' volunteer work for charity. The variables give the name of the school, the year of the fundraiser, the first name of each student, the amount of money that each student raised, and the number of hours that each student volunteered. The RETA IN statement and two sum statements compute the minimum and maximum values of Year . The CALL SYMPUT statements store these values in the macro variables FIRST_YEAR and LAST_YEAR. A DATA step creates this data set.

 data Charity;  input School $ 1-7 Year 9-12 Name $ 14-20 moneyRaised 22-26     hoursVolunteered 28-29;     format moneyRaised dollar8.2;     format hoursVolunteered f3.0;     format Year yrFmt.;     format School schFmt.;     label School = "Schools";     label Year = "Years";     retain yearmin yearmax;     yearmin=min(yearmin,year);     yearmax=max(yearmax,year);     call symput('first_year',put(yearmin,4.));     call symput('last_year',put(yearmax,4.));  datalines;  Monroe   1992 Allison 31.65 19  Monroe   1992 Barry   23.76 16  Monroe   1992 Candace 21.11  5  Monroe   1992 Danny    6.89 23  Monroe   1992 Edward  53.76 31  Monroe   1992 Fiona   48.55 13  Monroe   1992 Gert    24.00 16  Monroe   1992 Harold  27.55 17  Monroe   1992 Ima     15.98  9  Monroe   1992 Jack    20.00 23  Monroe   1992 Katie   22.11  2  Monroe   1992 Lisa    18.34 17  Monroe   1992 Tonya   55.16 40  Monroe   1992 Max     26.77 34  Monroe   1992 Ned     28.43 22  Monroe   1992 Opal    32.66 14  Monroe   1993 Patsy   18.33 18  Monroe   1993 Quentin 16.89 15  Monroe   1993 Randall 12.98 17  Monroe   1993 Sam     15.88  5  Monroe   1993 Tyra    21.88 23  Monroe   1993 Myrtle  47.33 26  Monroe   1993 Frank   41.11 22  Monroe   1993 Cameron 65.44 14  Monroe   1993 Vern    17.89 11  Monroe   1993 Wendell 23.00 10  Monroe   1993 Bob     26.88  6  Monroe   1993 Leah    28.99 23  Monroe   1994 Becky   30.33 26  Monroe   1994 Sally   35.75 27  Monroe   1994 Edgar   27.11 12  Monroe   1994 Dawson  17.24 16  Monroe   1994 Lou      5.12 16  Monroe   1994 Damien  18.74 17  Monroe   1994 Mona    27.43  7  Monroe   1994 Della   56.78 15  Monroe   1994 Monique 29.88 19  Monroe   1994 Carl    31.12 25  Monroe   1994 Reba    35.16 22  Monroe   1994 Dax     27.65 23  Monroe   1994 Gary    23.11 15  Monroe   1994 Suzie   26.65 11  Monroe   1994 Benito  47.44 18  Monroe   1994 Thomas  21.99 23  Monroe   1994 Annie   24.99 27  Monroe   1994 Paul    27.98 22  Monroe   1994 Alex    24.00 16  Monroe   1994 Lauren  15.00 17  Monroe   1994 Julia   12.98 15  Monroe   1994 Keith   11.89 19  Monroe   1994 Jackie  26.88 22  Monroe   1994 Pablo   13.98 28  Monroe   1994 L.T.    56.87 33  Monroe   1994 Willard 78.65 24  Monroe   1994 Kathy   32.88 11  Monroe   1994 Abby    35.88 10  Kennedy  1992 Arturo  34.98 14  Kennedy  1992 Grace   27.55 25  Kennedy  1992 Winston 23.88 22  Kennedy  1992 Vince   12.88 21  Kennedy  1992 Claude  15.62  5  Kennedy  1992 Mary    28.99 34  Kennedy  1992 Abner   25.89 22  Kennedy  1992 Jay     35.89 35  Kennedy  1992 Alicia  28.77 26  Kennedy  1992 Freddy  29.00 27  Kennedy  1992 Eloise  31.67 25  Kennedy  1992 Jenny   43.89 22  Kennedy  1992 Thelma  52.63 21  Kennedy  1992 Tina    19.67 21  Kennedy  1992 Eric    24.89 12  Kennedy  1993 Bubba   37.88 12  Kennedy  1993 G.L.    25.89 21  Kennedy  1993 Bert    28.89 21  Kennedy  1993 Clay    26.44 21  Kennedy  1993 Leeann  27.17 17  Kennedy  1993 Georgia 38.90 11  Kennedy  1993 Bill    42.23 25  Kennedy  1993 Holly   18.67 27  Kennedy  1993 Benny   19.09 25  Kennedy  1993 Cammie  28.77 28  Kennedy  1993 Amy     27.08 31  Kennedy  1993 Doris   22.22 24  Kennedy  1993 Robbie  19.80 24  Kennedy  1993 Ted     27.07 25  Kennedy  1993 Sarah   24.44 12  Kennedy  1993 Megan   28.89 11  Kennedy  1993 Jeff    31.11 12  Kennedy  1993 Taz     30.55 11  Kennedy  1993 George  27.56 11  Kennedy  1993 Heather 38.67 15  Kennedy  1994 Nancy   29.90 26  Kennedy  1994 Rusty   30.55 28  Kennedy  1994 Mimi    37.67 22  Kennedy  1994 J.C.    23.33 27  Kennedy  1994 Clark   27.90 25  Kennedy  1994 Rudy    27.78 23  Kennedy  1994 Samuel  34.44 18  Kennedy  1994 Forrest 28.89 26  Kennedy  1994 Luther  72.22 24  Kennedy  1994 Trey     6.78 18  Kennedy  1994 Albert  23.33 19  Kennedy  1994 Che-Min 26.66 33  Kennedy  1994 Preston 32.22 23  Kennedy  1994 Larry   40.00 26  Kennedy  1994 Anton   35.99 28  Kennedy  1994 Sid     27.45 25  Kennedy  1994 Will    28.88 21  Kennedy  1994 Morty   34.44 25  ;

Compute the descriptive statistics, and specify the options and subgroups for analysis. This PROC MEANS step analyzes the data for the one-way combination of the class variables and across all observations. It creates an output data set that includes variables for the total and average amount of money raised. The data set also includes new variables for the top three amounts of money raised, the names of the three students who raised the money, the years when the students raised the money, and the schools that the students attended.

 proc means data=Charity descendTypes charType noprint;     class School Year;     var moneyRaised;     types () School year;     output out=top3list sum= mean=        idgroup ( max(moneyRaised) out[3](moneyRaised name school year)= )        / autoname;     run;

Print the report. This PROC PRINT step generates traditional Listing output of the output data set that PROC MEANS created.

 proc print data=top3list nobs;     title 'Simple PROC PRINT of the Output Data Set';  run;

Listing Output from PROC PRINT

Listing 10.7: PROC PRINT Listing Output from PROC MEANS

 Simple PROC PRINT of the Output Data Set                 1                                  money    money                                Raised_  Raised_   money     money    money  School   Year _TYPE_ _FREQ_       Sum     Mean Raised_1  Raised_2 Raised_3  Kennedy   All   10      53   75.95   .73   .22    .63   .89  Monroe    All   10      56   16.80   .87   .65    .44   .87  All      1992   01      31    2.92   .80   .16    .76   .63  All      1993   01      32    7.92   .37   .44    .33   .23  All      1994   01      46   91.91   .26   .65    .22   .87  All       All   00     109   92.75   .29   .65    .22   .44  Name_1   Name_2   Name_3  School_1 School_2 School_3 Year_1 Year_2 Year_3  Luther   Thelma   Jenny   Kennedy   Kennedy Kennedy   1994   1992   1992  Willard  Cameron  L.T.    Monroe    Monroe  Monroe    1994   1993   1994  Tonya    Edward   Thelma  Monroe    Monroe  Kennedy   1992   1992   1992  Cameron  Myrtle   Bill    Monroe    Monroe  Kennedy   1993   1993   1993  Willard  Luther   L.T.    Monroe    Kennedy Monroe    1994   1994   1994  Willard  Luther   Cameron Monroe    Kennedy Monroe    1994   1994   1993

Program 2: Building a Custom Table Definition for the TopN Report

Set the SAS system options. The OPTIONS statement controls several aspects of the Listing output. None of these options affects the HTML output.

 options nodate pageno=1 pagesize=60 linesize=72;

Create the HTML output and specify the name of the HTML file. The ODS HTML statement opens the HTML destination and creates HTML output. It sends all output objects to the external file topn-body.htm in the current directory. Some browsers require an extension of .htm or .html on the filename.

 ods html body='topn-body.htm';

Create the table definition means.topn The DEFINE statement creates the table definition means.topn in the first template store in the path for which you have write access. By default, this template store is SASUSER.TEMPLAT.

 proc template;     define table means.topn;

Specify the symbols that reference three macro variables. The MVAR statement defines three symbols that reference macro variables. ODS will use the values of these variables as strings. References to the macro variables are resolved when ODS binds the definition and the data component to produce an output object. FIRST_YEAR and LAST_YEAR will contain the values of the first and last years for which there are data. Their values are assigned by the SYMPUT statements in the DATA step. SYSDATE9 is an automatic macro variable whose value is always available.

 mvar first_year last_year sysdate9;

Specify the column names and the order in which they appear in the report. The COLUMN statement declares these variables as columns in the table and specifies their order in the table. If a column name appears in parentheses, then PROC TEMPLATE stacks the values of all variables that use that column definition one below the other in the output object. Variables are assigned a column definition in the DATA step that appears later in the program.

 column class sum mean (raised) (name) (school) (year);

Specify three customized changes to the table definition. These three table attributes affect the presentation of the output object in the Listing output. They have no effect on its presentation in the HTML output. DOUBLE_SPACE= double spaces the rows of the output object. OVERLINE= and UNDERLINE= draw a continuous line before the first row of the table and after the last row of the table.

 double_space=on;  overline=on;  underline=on;

Specify the two table headers and the order in which they appear in the report. The HEADER statement declares table_header_1 and table_header_2 as headers in the table and specifies the order in which the headers appear in the output object.

 header table_header_1 table_header_2;

Create the table element table_header_1 . The DEFINE statement and its substatement and attribute define table_header_1 . The TEXT statement specifies the text of the header. The STYLE= attribute alters the style element that displays the table header. The style element header is defined in the default style definition, styles.default . (For information on viewing a style definition, see

'What Style Definitions Are Shipped with SAS Software?' on page 30.) In this case, the STYLE= attribute specifies a large font size . All other attributes that are included in header remain in effect. This attribute affects only the HTML output.

The END statement ends the header definition.

 define table_header_1;     text "Top Three Fund Raisers";     style=header{font_size=6};  end;

Create the table element table_header_2 . The DEFINE statement and its substatement and attribute define table_header_2 . The TEXT statement uses text and the macro variables FIRST_YEAR and LAST_YEAR to specify the contents of the header. When ODS binds the data component to the table definition (in the DATA step that follows ), it will resolve the values of the macro variables FIRST_YEAR and LAST_YEAR. The table definition itself contains references to the macro variables.

The SPACE= attribute inserts a blank line after the header (in the Listing output only).

The END statement ends the header definition.

 define table_header_2;      text "from " first_year " to " last_year;      space=1;  end;

Create the table element table_footer . The DEFINE statement and its substatement and attribute define table_footer . The FOOTER argument declares table_footer as a footer. (Compare this approach with the creation of the headers. You could use a FOOTER statement instead of the FOOTER argument in the DEFINE statement.)

The TEXT statement specifies the text of the footer. When ODS binds the data component to the table definition (in the DATA step that follows), it will resolve the value of the macro variable SYSDATE9. The table definition itself contains a reference to the macro variable. The SPLIT= attribute specifies the asterisk as the split character. This prevents the header from splitting at the open parenthesis. If no split character is specified, then ODS interprets the nonalphabetic, leading character as the split character (see the discussion of text-specification(s) in 'TEXT Statement' on page 409.) Alternatively, you can place a space character before the open parenthesis.

The STYLE= attribute alters the style element that displays the table footer. The style element header is defined in the default style definition, styles.default . (For information on viewing a style definition, see

'Viewing the Contents of a Style Definition' on page 319.) In this case, the STYLE= attribute specifies a small font size. All other attributes that are included in footer remain in effect. This attribute affects only the HTML output.

The END statement ends the footer definition.

 define footer table_footer;     text "(report generated on " sysdate9 ")";     split="*";     style=header{font_size=2};  end;

Create the column class . The DEFINE statement and its attributes create the column definition class . (The COLUMN statement earlier in the program declared class as a column.)

GENERIC= specifies that multiple variables can use the same column definition. GENERIC= is not specific to a destination.

ID= specifies that this column should be repeated on every data panel if the report uses multiple data panels. ID= affects only the Listing output.

VJUST= specifies that the text appear at the top of the HTML table cell that it is in. VJUST= affects only the HTML output.

STYLE= specifies that the column uses the DATA table element. This table element is defined in the default style definition, which is the style definition that is being used. STYLE= affects only the HTML output.

The END statement ends the definition.

Notice that, unlike subsequent column definitions, this column definition does not include a header. This is because the same header is not appropriate for all the variables that use this column definition. Because there is no header specified here or in the FILE statement, the header comes from the label that was assigned to the variable in the DATA step.

 define class;     generic=on;     id=on;     vjust=top;     style=data;  end;

Create six additional columns. Each of these DEFINE statements and its attributes creates a column definition. GENERIC= specifies that multiple variables can use a column definition (although in the case of sum and mean , only one variable uses the definition). HEADER= specifies the text for the column header. VJUST= specifies that the text appear at the top of the HTML table cell that it is in. The END statement ends the definition.

 define sum;     generic=on;     header="Total Dollars Raised";     vjust=top;  end;  define mean;     generic=on;     header="Average Dollars per Student";     vjust=top;  end;  define raised;     generic=on;     header="Individual Dollars";  end;  define name;     generic=on;     header="Student";  end;  define school;     generic=on;     header="School";  end;  define year;     generic=on;     header="Year";  end;

End the table definition. This END statement ends the table definition. The RUN statement ends the PROC TEMPLATE step.

 end;  run;

Create the data component. This DATA step does not create a data set. Instead, it creates a data component and, eventually, an output object. The SET statement reads the data set TOP3LIST that was created with PROC MEANS.

 data _null_;     set top3list;

Route the DATA step results to ODS and use the means.topn table definition. The combination of the fileref PRINT and the ODS option in the FILE statement routes the results of the DATA step to ODS. (For more information on using the DATA step with ODS, see Chapter 3, 'Output Delivery System and the DATA Step,' on page 39.) The TEMPLATE= suboption tells ODS to use the table definition named means.topn , which was previously created with PROC TEMPLATE.

 file print ods = (     template='means.topn'

Specify the column definition to use for each variable. The COLUMNS= suboption places DATA step variables into columns that are defined in the table definition. For example, the first column-specification specifies that the first column of the output object contains the values of the variable SCHOOL and that it uses the column definition named class . GENERIC= must be set to ON in both the table definition and each column assignment in order for multiple variables to use the same column definition.

 columns=(        class=school(generic=on)        class=year(generic=on)        sum=moneyRaised_sum(generic=on)        mean=moneyRaised_mean(generic=on)        raised=moneyRaised_1(generic=on)        raised=moneyRaised_2(generic=on)        raised=moneyRaised_3(generic=on)        name=name_1(generic=on)        name=name_2(generic=on)        name=name_3(generic=on)        school=school_1(generic=on)        school=school_2(generic=on)        school=school_3(generic=on)        year=year_1(generic=on)        year=year_2(generic=on)        year=year_3(generic=on)        )        );

Write the data values to the data component. The _ODS_ option and the PUT statement write the data values for all columns to the data component.

 put _ods_;  run;

Stop the creation of HTML output. The ODS HTML statement closes the HTML destination and all the files that are associated with it. You must close the destination before you can view the output with a browser.

 ods html close;

Listing Output for the TopN Report

Compare this customized output to the PROC PRINT listing output in Output 10.7.

Listing 10.8: Using a Customized Table to Produce Listing Output

 Simple PROC PRINT of the Output Data Set                      1                                 Top Three Fund Raisers                                        from  to                                  Average                           Total  Dollars                         Dollars      per    Individual    Schools    Years      Raised  Student       Dollars    Student  School      Year    ____________________________________________________________________________________    Kennedy      All    75.95   .73        .22    Luther   Kennedy     1994                                                 .63    Thelma   Kennedy     1992                                                 .89    Jenny    Kennedy     1992    Monroe       All    06.80    .69       .65    Willard  Monroe      1994                                                 .44    Cameron  Monroe      1993                                                 .87    L.T.     Monroe      1994    All         1992     2.92    .48       .16    Tonya    Monroe      1992                                                 .76    Edward   Monroe      1992                                                 .63    Thelma   Kennedy     1992    All         1993     7.92    .37       .44    Cameron  Monroe      1993                                                 .33    Myrtle   Monroe      1993                                                 .23    Bill     Kennedy     1993       All         1994    91.91    .26       .65    Willard  Monroe      1994                                                 .22    Luther   Kennedy     1994                                                 .87    L.T.     Monroe      1994    All          All    82.75    .20       .65    Willard  Monroe      1994                                                 .22    Luther   Kennedy     1994                                                 .44    Cameron  Monroe      1993  _______________________________________________________________________________________                           (report generated on 30JUN2003)

HTML Output: Using a Customized Table for the TopN Report

Display 10.9: HTML Output for the TopN Report (Viewed with Microsoft Internet Explorer)

Example 4: Changing a Column without Redefining the Table Definition

PROC TEMPLATE features:

DEFINE TABLE statement
Table attributes:

Other ODS features:

ODS HTML statement

Program Description

Program

 proc template;     define table Base.Summary;        notes "Summary table for MEANS and SUMMARY";        dynamic clmpct one_var_name one_var_label one_var;        column class nobs id type ways (varname) (label) (min) (max) (range) (n           ) (nmiss) (sumwgt) (sum) (mean) (uss) (css) (var) (stddev) (cv) (           stderr) (t) (probt) (lclm) (uclm) (skew) (kurt) (median) (mode) (q1)           (q3) (qrange) (p1) (p5) (p10) (p25) (p50) (p75) (p90) (p95) (p99);        define nobs;          style={foreground=green};           id;        end;     end;  run;  ods html file="tmep.html";  proc summary data=sashelp.class print;  class age;  run;  ods html close;

Example 5: Setting the Style Element for Cells Based on Their Values

PROC TEMPLATE features:

DEFINE TABLE statement
- NMVAR statement
- NOTES statement
- TRANSLATE-INTO statement

DEFINE COLUMN statement
- BLANK_DUPS= attribute
- CELLSTYLE-AS statement
- GENERIC= attribute

Other ODS features:

ODS HTML statement
FILE statement with ODS= option
PUT statement with _ODS_ argument

Data set: GRAIN_PRODUCTION'' on page 97

Format: $CNTRY.'' on page 97

Program Description

This example creates a template that uses different colors and font attributes for the text inside cells, depending on their values.

Program

Set the SAS system options. The OPTIONS statement controls several aspects of the Listing output. None of these options affects the HTML output. The TITLE statement specifies a title.

 options nodate pageno=1 pagesize=60 linesize=72;  title 'Leading Grain Producers';

Create the table definition shared.cellstyle . The DEFINE statement creates the table definition shared.cellstyle in the first template store in the path that is available to write to. By default, this template store is SASUSER.TEMPLAT.

 proc template;     define table shared.cellstyle;

Specify that missing values show the text 'No data' in the report. The TRANSLATE-INTO statement translates missing values (.) into the string No data .

 translate _val_=. into 'No data';

Store the information about the table in the table definition. The NOTES statement provides information about the table. NOTES statements remain a part of the compiled table definition whereas SAS comments do not.

 notes "NMVAR defines symbols that will be used to determine the colors  of the cells.";

Specify the symbols that reference three macro variables. The NMVAR statement defines three symbols that reference macro variables. ODS will convert the variable's values to numbers (stored as doubles) before using them. References to the macro variables are resolved when ODS binds the definition and the data component to produce an output object. The text inside quotation marks provides information about the symbols. This information becomes a part of the compiled table definition whereas SAS comments do not.

LOW, MEDIUM, and HIGH will contain the values to use as the determinants of the style element that is used to display the cell. The values are provided just before the DATA step that produces the report.

 nmvar low 'Use default style.'        medium 'Use yellow foreground and bold font weight'        high 'Use red foreground and a bold, italic font.';

Control the repetition of values that do not change from one row to the next row. The CLASSLEVELS= attribute suppresses the display of the value in a column that is marked with BLANK_DUPS=ON if the value changes in a previous column that is also marked with BLANK_DUPS=ON. Because BLANK_DUPS= is set in a generic column, you should set this attribute as well.

 classlevels=on;

Create the column char_var . The DEFINE statement and its attributes create the column definition char_var . GENERIC= specifies that multiple variables can use the same column definition. BLANK_DUPS= suppresses the display of the value in the column if it does not change from one row to the next (and, because CLASSLEVELS=ON for the table, if no value changes in a preceding column that is marked with BLANK_DUPS=ON changes).

The END statement ends the definition.

 define column char_var;     generic=on;     blank_dups=on;  end;

Create the column num_var . The DEFINE statement and its attributes create the column definition num_var . GENERIC= specifies that multiple variables can use the same column definition.

 define column num_var;     generic=on;

Align the values in the column without regard to the format field. JUSTIFY= justifies the values in the column without regard to the format field. For numeric variables, the default justification is RIGHT, so even the translated character value No data that is used for missing values is right-justified. Without JUSTIFY=ON in this column definition, the value No data is formatted as a character variable (left-justified) within a format field that has the same width as the column.

 justify=on;

Specify which style element and style attributes to use for different values in the column. The CELLSTYLE-AS statement specifies the style element and style attributes to use for different values in this column. If a value is less than or equal to the value of the variable LOW, the cell uses the unaltered Data style element. If a value is greater than LOW but less than or equal to the value of MEDIUM, the cell uses the style element Data with a foreground color of green and an italic font. Similarly, other values use a foreground color of yellow or red and combinations of a bold font weight and an italic font style. The CELLSTYLE-AS statement affects only the HTML destination.

The END statement ends the column definition.

 cellstyle _val_ <= low as data,                  _val_ <= medium as data                           {foreground=green font_style=italic},                  _val_ <= high as data                           {foreground=yellow font_weight=bold},                      1 as data                           {foreground=red font_style=italic                            font_weight=bold};  end;

End the table definition. This END statement ends the table definition. The RUN statement ends the PROC TEMPLATE step.

 end;  run;

Create the HTML output and specify name of the HTML file. The ODS HTML statement opens the HTML destination and creates HTML output. It sends all output objects to the external file cellstyle-body.htm in the current directory. Some browsers require an extension of .htm or .html on the filename.

 ods html body='cellstyle-body.htm';

Assign values to three macro variables. The %LET statements assign values to the macro variables LOW, MEDIUM, and HIGH.

 %let low=10000;  %let medium=50000;  %let high=100000;

Create the data component. This DATA step does not create a data set. Instead, it creates a data component, and, eventually, an output object. The SET statement reads the data set GRAIN_PRODUCTION.

 data _null_;     set grain_production;

Route the DATA step results to ODS and use the shared.cellstyle table definition. The combination of the fileref PRINT and the ODS option in the FILE statement routes the results of the DATA step to ODS. (For more information on using the DATA step with ODS, see Chapter 3, 'Output Delivery System and the DATA Step,' on page 39.) The TEMPLATE= suboption tells ODS to use the table definition named shared.cellstyle , which was previously created with PROC TEMPLATE.

 file print ods=(       template='shared.cellstyle'

Specify the column definition to use for each variable. The COLUMNS= suboption places DATA step variables into columns that are defined in the table definition. For example, the first column-specification specifies that the first column of the output object contains the values of the variable YEAR and that it uses the column definition named char_var . GENERIC= must be set to ON, both in the table definition and in each column assignment, in order for multiple variables to use the same column definition.

 columns=(     char_var=year(generic=on)     char_var=country(generic=on format=$cntry.)     char_var=type(generic=on)     num_var=kilotons(generic=on format=comma12.)     )  );

Write the data values to the data component. The _ODS_ option and the PUT statement write the data values for all columns to the data component.

 put _ods_;  run;

 ods html close;

Only the table customizations appear in the Listing output. Table customizations include the suppression of values that do not change from one row to the next and the translation of missing values to No data . The style customizations that are specified in the CELLSTYLE-AS statement do not appear in the Listing output.

Listing Output of a Customized Table

Output 10.9: Listing Output

 Leading Grain Producers                             1  Year     Country            Type           Kilotons  1995     Brazil             Corn             36,276                              Rice             11,236                              Wheat             1,516           China              Corn            112,331                              Rice            185,226                              Wheat           102,207           India              Corn              9,800                              Rice            122,372                              Wheat            63,007           Indonesia          Corn              8,223                              Rice             49,860                              Wheat           No data           United States      Corn            187,300                              Rice              7,888                              Wheat            59,494  1996     Brazil             Corn             31,975                              Rice             10,035                              Wheat             3,302           China              Corn            119,350                              Rice            190,100                              Wheat           109,000           India              Corn              8,660                              Rice            120,012                              Wheat            62,620           Indonesia          Corn              8,925                              Rice             51,165                              Wheat           No data           United States      Corn            236,064                              Rice              7,771                              Wheat            62,099

HTML Output of a Customized Table

Both the table customizations and the style customizations appear in the HTML output. Table customizations include the suppression of values that do not change from one row to the next, and the translation of missing values to No data . The style customizations include the colors and font styles that are specified in the CELLSTYLE-AS statement.

Display 10.10: HTML Output (Viewed with Microsoft Internet Explorer)

Example 6: Setting the Style Element for a Specific Column, Row, and Cell

PROC TEMPLATE features:

DEFINE STYLE statement
- REPLACE statement

DEFINE TABLE statement

CELLSYTLE-AS statement

DEFINE COLUMN statement
- DEFINE HEADER statement
  - TEXT statement

DEFINE HEADER statement
- TEXT statement

Other ODS features:

FILE statement with ODS= option
ODS HTML statement
- STYLE= option

ODS PDF statement
- STYLE= option

PUT statement with _ODS_ argument
ODS TRACE statement

Program Description

This example combines a customized style definition with a customized table definition to produce output with a checkerboard pattern of table cells.

Program

Create a new style definition Greenbar . The PROC TEMPLATE statement starts the TEMPLATE procedure. The DEFINE STYLE statement creates a new style definition greenbar .

 proc template;     define style greenbar;

Specify the parent style definition from which the greenbar style definition inherits its attributes. The PARENT= attribute specifies the style definition from which the greenbar definition inherits its style elements and attributes. All the style elements and their attributes that are specified in the parent's definition are used in the current definition unless the current definition overrides them.

 parent=styles.printer;

Change the colors used in the headers and footers. The REPLACE statement adds a style element to the greenbar style definition from the parent style definition, but the background is light green and the foreground is black.

 replace headersandfooters from cell /        background=light green        foreground=black     ;

End the style definition. The END statement ends the style definition. The RUN statement executes the PROC TEMPLATE step.

 end;  run;

Create the HTML and PDF output and specify the style definition that you want to use for the output. The ODS HTML statement opens the HTML destination and creates HTML output. It sends all output objects to the file greenbar.html in the current directory. The STYLE= option tells ODS to use greenbar as the style definition when it formats the output.

The ODS PDF statement opens the PDF destination and creates PDF output. It sends all output objects to the file greenbar.pdf in the current directory. The STYLE= option tells ODS to use greenbar as the style definition when it formats the output.

 ods html body="greenbar.html" style=greenbar;  ods pdf file="greenbar.pdf" style=greenbar;

Create the table definition Checkerboard . The DEFINE statement creates the table definition Checkerboard in the first template store in the path that is available to write to. By default, this template store is SASUSER.TEMPLAT.

 proc template;     define table Checkerboard;

Specify which style element and style attributes to use for different cells.

The CELLSTYLE-AS statement specifies the style element and style attributes to use for cells in each of the rows and columns. The CELLSTYLE-AS statement creates the checkerboard effect in the output. If both the row and column are odd numbered, then the cell is magenta in color. Similarly, if both the row and column are even numbered, then the cell is magenta in color. The CELLSTYLE-AS statement has no effect on the LISTING destination because it is changing style elements and style attributes which have no effect in listing output.

 cellstyle mod(_row_,2) && mod(_col_,2) as data{background=magenta},                  not(mod(_row_,2)) && not(mod(_col_,2)) as data{background=magenta},                  1 as data;

Create the header definition top . The DEFINE HEADER statement defines the table header top .

The TEXT statement specifies the text of the header Checkerboard Table Definition .

The END statement ends the header definition.

 define header top;     text 'Checkerboard Table Definition';  end;

Create the column definition name . The DEFINE COLUMN statement creates the column definition name .

The DEFINE HEADER statement creates the header definition bar .

The TEXT statement specifies the text to use in the header. _LABEL_ is a dynamic variable that references a value that the data component supplies from the procedure or DATA step, in this example the variable's label.

The first END statement ends the header definition.

The HEADER statement declares bar as the header in the table.

The second END statement ends the column definition.

 define column name;     define header bar;        text "begin " _label_ " end";     end;     header=bar;  end;

Create the column definition gender . The DEFINE COLUMN statement creates the column definition gender .

The DATANAME= column attribute specifies the name of the column sex in the data component to associate with the column definition gender .

The DEFINE HEADER statement creates the header bar .

The TEXT statement specifies the text Gender to use in the header.

The first END statement ends the header definition.

The HEADER statement declares bar as the header in the table.

The second END statement ends the column definition.

 define column gender;     dataname=sex;     define header bar;        text "Gender";     end;     header=bar;  end;

Create three column definitions: age , height , and weight . The three DEFINE COLUMN statements create the column definitions age , weight , and height .

The END statement ends the column definition.

 define column age;  end;  define column height;  end;  define column weight;  end;

End the table definition. The END statement ends the table definition. The RUN statement executes the TEMPLATE procedure.

 end;  run;

Create the data component. This DATA step does not create a data set. Instead, it creates a data component that is used to produce an output object.

The SET statement reads the data set SASHELP.CLASS .

 data _null_;     set sashelp.class;

Route the DATA step results to ODS and use the Checkerboard table definition. The combination of the fileref PRINT and the ODS option in the FILE statement routes the results of the DATA step to ODS. (For more information about using the DATA step with ODS, see Chapter 3, 'Output Delivery System and the DATA Step,' on page 39.) The TEMPLATE= suboption tells ODS to use the table definition named Checkerboard .

 file print ods=(template='Checkerboard');    put _ods_;  run;

Stop the creation of HTML and PDF output. The ODS HTML statement closes the HTML destination and all the files that are associated with it. The ODS PDf statement closes the PDF destination and all the files that are associated with it. You must close the destinations before you can view the output.

 ods html close;  ods pdf close;

Display 10.11: HTML Output (Viewed with Internet Explorer 6.0)

Display 10.12: PDF Output (Viewed with Acrobat Reader 5.0)